Читать книгу Book Wars - John B. Thompson - Страница 14

The issues are more complicated than they appear at first sight, however. While the S-curve gives a neat picture of the overall trend, it is misleading because it collapses all the different kinds of books into one average number. We shouldn’t assume that different kinds of books perform in the same way – they don’t. The extent to which books have migrated from print to digital formats has varied enormously from one kind of book to another. We can see this by looking at some sales data from a large US trade publisher that I’ll call ‘Olympic’. Table 1.5 and figures 1.6a and 1.6b give a breakdown of ebook sales as a percentage of total sales at Olympic from 2006 to 2016. All data are based on net units and net sales – that is, sales net of any returns. We should not assume that these data are representative of the industry as a whole, or that the experiences of all trade publishers will have been identical to this one – the data from each publishing house will be unique and will reflect to some extent the specific titles they’ve published. But Olympic is a mainstream trade publisher with a large and varied list, and the occasional truly exceptional title has been stripped out of the data to minimize the distorting effect of outliers. So while the experiences of each publisher will be unique, it is unlikely that the sales patterns of other major trade houses will differ radically from the patterns experienced by Olympic.

Table 1.5 Ebooks as a percentage of total sales at Olympic, units and dollars

	Ebooks units	Ebooks $
2006	0.1	0.1
2007	0.1	0.1
2008	0.5	0.5
2009	1.9	2.6
2010	6.2	8.6
2011	16.4	19.5
2012	22.2	25.9
2013	20.7	23.8
2014	19.8	23.7
2015	19.7	22.6
2016	16.4	17.1

Figure 1.6a Ebooks as a percentage of total sales at Olympic, units and dollars

Figure 1.6b Ebooks as a percentage of total sales at Olympic, units and dollars

There is another important qualification to make about these data: they are for the period 2006–16 only and we cannot extrapolate, on the basis of these data, to the sales patterns for 2017 and subsequent years. Future patterns could change. I’ll return to this issue below. But, for now, let’s focus on what actually happened in the crucial decade from 2006 to 2016.

Table 1.5 and figure 1.6a show all ebooks as a percentage of Olympic’s total sales by both units and revenue. We see that Olympic’s ebook sales were negligible in 2006 and 2007 but they began to grow rapidly from 2008 on, reaching a peak in 2012, when ebook sales accounted for just under 26% of Olympic’s total revenue. From that point on, ebook sales began to decline as a proportion of total sales, falling to below 23% in 2015 and then down to 17% in 2016. The pattern is very similar for both units and revenue, as one would expect. The levelling off in ebook sales is more vividly displayed when we change the scale of the y-axis on the graph, as in figure 1.6b: here, again, we see that the pattern of ebook sales at Olympic displays the classic technology S-curve.

However, looking at all ebooks as a percentage of total sales gives us a very partial view of what has happened because it masks the variations between different categories of books. In the early 2000s, before ebooks began to take off, many commentators assumed that when the ebook revolution began, it would be driven primarily by businessmen who wanted to carry business books with them on their business trips, reading at airports and on planes: it was adult nonfiction, and especially business books and ‘big idea’ books, that would, they thought, spearhead the ebook revolution. Were they right? Is that what actually happened?

Table 1.6 and figure 1.7 break down ebooks into three broad categories: adult fiction, adult nonfiction and juvenile (where juvenile includes all children’s books as well as young adult). The figure shows ebooks as a percentage of Olympic’s total sales, by both revenue and units, in each of these broad categories. (As with figure 1.6b, the y-axis has been adjusted to display the S-curve.) It is immediately clear that the category where the biggest change has occurred is not nonfiction but, rather, adult fiction: in terms of revenue, ebooks as a percentage of total revenue in adult fiction increased from 1.0% in 2008 to 43.4% in 2014, before falling back to 37.4% in 2015 and then rebounding slightly to 38.9% in 2016. This contrasts sharply with adult nonfiction, where ebooks as a percentage of total revenue rose from 0.4% in 2008 to 16.6% in 2015, before falling back to 13.2% in 2016 – remaining well below 20% throughout this period. Juvenile lagged even further behind: here ebooks as a percentage of total revenue increased from 0.1% in 2008 to 12.2% in 2014, before falling back to 7.4% in 2015 and 6% in 2016.⁵ Each broad category displays an S-curve but the shape of the curve is different in each case. Adult fiction nearly reaches 45% before it levels off and starts to decline, adult nonfiction levels off at around 15% and, in the case of juvenile, the ebooks peak at around 12% and then fall off. Both adult fiction and juvenile show a sharp downturn in 2015, while adult nonfiction continues to grow very modestly before falling off in 2016.

Table 1.6 Ebooks as a percentage of total sales by broad category at Olympic, units and dollars

	Adult Fiction units	Adult Fiction $	Adult NF units	Adult NF $	Juv units	Juv $
2006	0.2	0.1	0.1	0.1	0	0
2007	0.2	0.2	0.1	0.1	0	0
2008	0.9	1	0.4	0.4	0.1	0.1
2009	4	4.7	1.5	1.8	0.2	0.4
2010	12.6	14.3	4.4	4.9	0.9	1.4
2011	29.1	30.4	12.2	12.6	3.5	5.3
2012	37.2	38.2	15.9	15	5	8.4
2013	40.9	40.2	16	15.3	5.9	9.2
2014	42.6	43.4	16.6	15.8	7.5	12.2
2015	40.6	37.4	18.2	16.6	4	7.4
2016	35	38.9	16.4	13.2	6	6

Figure 1.7 Ebooks as a percentage of total sales by broad category at Olympic, units and dollars

We are still working with very broad categories, however. Let’s drill down a little further and examine the patterns for different categories of books, using a selected number of standard BISAC subject headings.⁶ Figures 1.8 and 1.9 break down ebooks as a percentage of total sales by subject at Olympic (for the data on which these figures are based, see appendix 1). Figure 1.8 is net dollars and Figure 1.9 is net units. (Again, the y-axis has been adjusted to display the S-curves.) These graphs vividly display the enormous variation in the uptake of ebooks across different categories of books, and underscore how misleading it is to collapse all of these categories into the single category of ‘ebook’. We see the huge spectrum of trajectories here, with each category of book displaying its own distinctive S-curve. Each S-curve rises in its own unique way and begins to level off at a point and in a manner that is specific to that category. In some cases, the growth plateaus and then stabilizes more or less at that level; in some cases, it plateaus and then begins to decline; in some cases, the growth levels off and declines and then shoots back up; and in other cases, the growth never takes off at all. We also see a lot of movement up and down – the lines jump about, dips are followed by rises and rises are followed by dips as the numbers for each category fluctuate from one year to the next. There’s nothing too surprising about that: these graphs are based on sales figures from one large trade house which has a limited number of books in any one category in any one year, so one or two books selling strongly as ebooks (or other special circumstances, like the disposal or acquisition of an imprint) can produce an ebook spike or dip in that category. The sales figures from one publisher – even a large publisher like Olympic – will display idiosyncrasies of this kind and therefore cannot be taken as a proxy for the industry as a whole. But by focusing on the broad patterns and trends rather than the fluctuations from one year to the next, we can get a good sense of how different categories of books have performed over time.

Figure 1.8 Ebooks as a percentage of total sales by subject at Olympic, net dollars

Figure 1.9 Ebooks as a percentage of total sales by subject at Olympic, net units

As these data make clear, the top-performing category in terms of ebook uptake is not business books after all, it is romance fiction – this outperforms every other category by a significant margin. Here we see steep growth from 2008 to 2011, by which time ebooks were accounting for 44.2% of all Olympic’s sales of romance books. Ebook sales dipped the following year but then rose again, accounting for around 55% of all romance sales in 2013 and 2014. In 2015 they fell back to 45% but then rebounded in 2016, when they once again amounted to around 53% of all Olympic’s sales of romance books. Of all the different categories of books published by Olympic, romance is the one where ebooks have accounted for the highest proportion of overall sales – more than half – and it’s a category where ebook sales remain high despite the downturn in other categories.

At the other end of the scale, juvenile nonfiction has seen very low levels of ebook uptake. The line for this category is flat and hardly rises off the floor of the graph – ebooks accounted for only 2% of Olympic’s revenue in the category of juvenile nonfiction in 2015, rising slightly to 2.6% in 2016. Here we don’t see an S-curve because there has not yet been any perceptible take off in terms of ebook sales in this category of books: 97% of the revenue in 2016 was still being generated by printed books.

Between romance at the top and juvenile nonfiction at the bottom there is a huge range and variation in terms of ebook uptake – each category leaves its own distinctive footprint. But while the trajectories are all unique, the lines band together in certain groups. The top four lines all represent categories of fiction, and the top three lines are all genre fiction categories – romance at the top, followed by mystery and detective fiction, and then sci-fi and fantasy. General fiction is among this set of categories in which ebooks perform strongly, although the line for general fiction is below the genre fiction lines.⁷ Ebook sales in all four categories show a steep rise between 2008 and 2012, reaching levels that are much higher than with other categories of books. While romance plateaus at around 55% and then fluctuates after that, the other three fiction categories plateau at between 30 and 40%. Most of these categories display some modest decline after the peak period of 2012–14, though ebook sales of romance titles at Olympic experienced a new upsurge in 2016.

The next band of lines in the middle of the graphs all represent nonfiction categories – biography and autobiography, history, business and economics, family and relationships, health and fitness, religion, self-help. Once again, all of these lines rise steeply in the period between 2008 and 2011 and then begin to level off, though at lower levels than the fiction categories. Biography and autobiography and history continued to edge upwards after 2011, reaching 27% in 2015, and then fell sharply after that. Health and fitness reached 24% in 2015 and then began to fall. Other nonfiction categories like business and economics, family and relationships, religion and self-help reached plateaus of between 15 and 20%, and either stabilized at that level or began to fall off. So, between 2011 and 2015, all of these nonfiction categories appear to have plateaued somewhere between 16 and 27% – or, to put it more roughly, between 15 and 25%, with biography and autobiography and history at the top of this band.

At the bottom of the graphs is another set of categories where ebooks have so far failed to take off in any significant way and, by 2016, they still accounted for only a small proportion of overall revenue. Here we find cooking, where ebooks never rose above 5% of sales, juvenile nonfiction, where ebook sales never rose above 3% of sales, and juvenile fiction, where ebooks rose to 12.7% in 2014 and then fell back to 6% in 2016.⁸ Travel also belongs in this band: ebook sales in travel never took off at Olympic, generally remaining below 12% of total sales, and the spike in 2016 was an anomaly accounted for by particular circumstances at the time. Given that ebooks have not taken off in any significant way in these categories (excluding young adult), the growth lines since 2008 don’t display the pattern of the classic S-curve: they look more like flat lines with gentle inclines that tilt upwards (and occasionally downwards).